For many years, darkfield scanning methodologies have been used to scan patterned surfaces. Darkfield scanning makes use of light scattered or diffracted by the surface to characterize and examine features of the surface. As used herein, scattered light shall refer to both scattered light and diffracted light. FIG. 1 is a cross-section view of an illuminated surface used to illustrate aspects of darkfield scanning. An illumination source 101 projects a light beam I (also referred to herein as the incident beam) onto the surface 102 being examined. A portion of the incident beam I is reflected by the surface as the reflected beam R. If the surface 102 was perfectly reflective, the entire incident beam I would be reflected. However, most surfaces have a variety of characteristics which cause a portion of the light from an incident beam I to be scattered. Darkfield scanning makes use of this scattered light.
One particular surface feature that causes light scattering is referred to as a defect. The detection, quantification, and classification of defects is important in many areas. In particular, defect detection and analysis are important in semiconductor processing. Defects include, but are not limited to, pits, bumps, scratches, and a number of other features, which mar the surface 102. Thus, the light of an incident beam I is often subject to some degree of scattering. FIG. 1 illustrates a typical incident beam I having a light scattering pattern schematically depicted by a plurality of scattered light rays 103, 104, 105, and 106, which are scattered by a surface defect 108. The depicted plurality of rays can represent a continuous angular distribution of light scattered and diffracted by the surface.
Most conventional darkfield inspection tools make use of a single discrete photosensor element (for example a photomultiplier tube (PMT)) to detect the light scattered from the inspection surface. Some designs use as many as three or four distinct and widely separated discrete photodetector elements. Such discrete photodetector element(s) are positioned so that they are not in the path of the specular (reflected) beam R. This results in a detection field where the background (the field) background or field. Hence, the name darkfield scanning.
In a typical inspection tool, an illumination source directs an incident light beam onto the surface being inspected (i.e., a workpiece that is commonly, but not exclusively, a semiconductor wafer). If the surface were perfectly reflective, all light would be reflected in the specular direction (R of FIG. 1). However, under most conditions, even the highest quality wafers (or other surfaces) have some degree of surface roughness which causes scattering of the incident light beam. Moreover, surface imperfections and other defects give rise to further scattering. It is this concept of light scattering by surface defects that forms the foundation of conventional darkfield inspection techniques used for defect detection.
FIGS. 2(a) and 2(b) depicts cross-section views of a surface being scanned using darkfield scanning. The surface 102 is illuminated by an incident beam I, a portion of which is reflected as reflected beam R. Other portions of the incident beam I are scattered. FIG. 2(a) depicts the scatter from the surface 102 in the absence of a defect. Since ordinary surfaces are not perfect the incident light is scattered at a number of different angles. This results in a three-dimensional angular light distribution that can be different for each wafer depending on surface characteristics (e.g., surface topography, thickness and type of materials used, the layered structure of the surface, and so on) and other factors. This three-dimensional angular light distribution is referred to herein as the ordinary scattering pattern 200 of the surface 102 being inspected. Because FIG. 2(a) is a two dimensional representation of a three dimensional reality, only one range of scattering angles is depicted for the ordinary scattering pattern 200. In actuality the scattering angles of the ordinary scattering pattern 200 extend into and out of the page.
FIG. 2(b) depicts the same surface 102 as depicted in FIG. 2(a) except that the surface has a defect D formed thereon. The presence of the defect D causes the scattering pattern to vary. The defect D scatters some light as, for example, scattered light rays S1, S2, and S3. Additionally, much of the light still falls within the scattering angles defined by the ordinary scattering pattern 200′. It is the measurement of this scattered light that enables the inspection tool to detect and characterize defects in an inspected surface 102.
As depicted in the simplified schematic depiction of FIG. 3(a), in some implementations of conventional darkfield inspection a wafer 300 is placed in a tool and a spiral inspection pattern 301 is performed. During such an inspection the light scattered from the surface of the wafer 300 is detected. The intensity (I) of the scattered light can be plotted over time (t) as depicted in FIG. 3(b). As is commonly the case, the intensity of the scatter increases when the incident beam illuminates a defect. Such a defect signal 302 is schematically depicted. And because in a spiral inspection pattern (as well as many other inspection patterns) time correlated to the position of the defect, the defect can be located and identified. However, due to the small size of the defect, the increase in scattered light intensity can be very slight (as shown by the slight increase in amplitude for the defect signal 302). Thus, one of the challenges in conventional darkfield inspections of this type is to enhance the signal-to-noise ratio (SNR) for such inspection increasing the reliability and sensitivity of such inspections. Thus, what is needed are improved methods and apparatus for receiving and processing defect signals generated using scattered light in inspection processes.